Solvent engineering: an effective tool to direct chemoselectivity in a lipase-catalyzed Michael addition

Article abstract of DOI:10.1016/j.tet.2008.10.103

A solvent engineering strategy was implemented in order to control the chemoselectivity in a lipase-catalyzed Michael addition reaction. This strategy was revealed as a high-effective tool for the selective synthesis of Michael adduct 3 or aminolysis product 4 from benzylamine 1 and methyl crotonate 2. Chemoselectivity of the enzymatic process was elucidated in terms of polarity of the medium, hence, adduct 3 was preferentially accumulated in hydrophobic medium, whereas in polar solvents the amide 4 was preferentially formed.

Full text of DOI:10.1016/j.tet.2008.10.103

Tetrahedron 65 (2009) 536–539
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Solvent engineering: an effective tool to direct chemoselectivity
in a lipase-catalyzed Michael addition
a
b
b
a
´
´
´
´
Jaime Priego , Claudia Ortız-Nava , Manuel Carrillo-Morales , Agustın Lopez-Munguıa ,
Jaime Escalante b , Edmundo Castillo
,
a,
*
*
a
´
´
´
Instituto de Biotecnologıa, Universidad Nacional Autonoma de Mexico, Apartado Postal 510-3, C.P. 62271 Cuernavaca, Morelos, Mexico
b Centro de Investigaciones Qu´ımicas, Universidad Auto´noma del Estado de Morelos, Av. Universidad No. 1001, C.P. 62209 Cuernavaca, Morelos, Mexico
a r t i c l e i n f o
a b s t r a c t
Article history:
A solvent engineering strategy was implemented in order to control the chemoselectivity in a lipase-
catalyzed Michael addition reaction. This strategy was revealed as a high-effective tool for the selective
synthesis of Michael adduct 3 or aminolysis product 4 from benzylamine 1 and methyl crotonate 2.
Chemoselectivity of the enzymatic process was elucidated in terms of polarity of the medium, hence,
adduct 3 was preferentially accumulated in hydrophobic medium, whereas in polar solvents the amide 4
was preferentially formed.
Received 22 October 2008
Accepted 28 October 2008
Available online 6 November 2008
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
reaction, alternating between a Michael addition and an aminolysis
product, it is important to point out the essential effect that sol-
In the last years, enzymes in organic media have become
a useful alternative in organic synthesis due to their chemo-, regio-,
and enantioselective properties.1 Particular attention has been paid
to lipase B from Candida antarctica (CaLB), because of its high sta-
bility and activity suitable for a broad spectrum of reactions, sub-
strates, and media. This biocatalyst represents an effective tool to
overcome significant limitations found in traditional organic
synthesis.2
vents can also have on the reaction thermodynamics.
2. Results and discussions
We have previously reported the use of solvent engineering
strategies in lipase-catalyzed reactions in order to manipulate the
Recently, novel and valuable applications of CaLB were reported:
Michael addition type reactions involving addition of secondary
amines to acrylonitrile3 or addition of imidazoles to acrylic
monomers in organic medium4 among other reactions. These ad-
Oxyanion
hole
CH3
O
Thr
40
O
O
H
O
N
N
H
H
ditions take place when a,b-unsaturated systems are used as the
Gln
electrophile moiety and amines as the nucleophile substrate. Some
evidences of the mechanism of this promiscuous activity of CaLB
point out that the oxyanion hole (Thr40 and Gln106) of the active site
stabilizes the negative charge of the transition state while the
His224-Asp187 pair facilitates proton transfer during the catalysis.
Consequently, it is proposed that solvents of low polarity may
induce interaction between the oxyanion hole and the carbonyl
oxygen in the catalytic intermediate complex, allowing the ability
of CaLB to carry out this reaction (Fig. 1).5
106
MeO
CH3
H
H
N
H
N
N
H
Although the mechanism described above may help to explain
the effect of reaction conditions on the selectivity of the enzymatic
O
O
His
224
Asp
187
* Corresponding authors. Tel.: þ52 55 56 22 76 09; fax: þ52 77 73 11 49 03.
E-mail addresses: jaime@ciq.uaem.mx (J. Escalante), edmundo@ibt.unam.mx
(E. Castillo).
Figure 1. Hypothetical catalytic addition of benzylamine to methyl crotonate in the
active site of CaLB. The carbonyl oxygen of the -unsaturated is bound to the oxy-
anion hole and the nucleophile is activated by His.5
a,b
0040-4020/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2008.10.103

J. Priego et al. / Tetrahedron 65 (2009) 536–539
537
thermodynamic activity of the reactants, allowing the accumula-
tion of a specific product in the reaction medium at equilibrium.6
We have also demonstrated that in inverse hydrolysis reactions,
polar products such as amides or polar esters are preferentially
accumulated in polar solvents while in hydrophobic solvents their
synthesis is quite disfavored.7 In fact, the higher the solubility of
a product in a given solvent the lower its thermodynamic activity
coefficient favoring the accumulation of the product at equilibrium
and vice versa.6
Based on these thermodynamic considerations, in this work we
make use of a solvent engineering strategy as an effective tool to
control chemoselectivity between a Michael addition and the
aminolysis product when benzylamine 1 and methyl crotonate 2
are exposed to CaLB as catalyst (Scheme 1).
In particular, for n-hexane,
a
non-polar solvent with an
ET(30)¼30.9 (entry 1), 95% of converted 1 corresponds to the Mi-
chael adduct product 3 while only 5% is transformed to 4. On the
other hand, in 2M2B, a polar solvent with an ET(30)¼41.0 (entry 7),
the proportion of converted 1 to 3 and 4 was 20 and 80%, re-
spectively. Actually, we have reported earlier7,10 that a lower ther-
modynamic activity of amides in 2M2B leads to its preferential
formation while in non-polar solvent a higher thermodynamic
activity results in lower yield. It is noteworthy that the product 1
obtained in n-hexane did not show significant specific optical ro-
tation [[
in 2M2B showed a specific optical rotation of 11.0 [[a]
20 11.0 (c 1.0,
D
a]
20 0.45 (c 1.0, CHCl3)], on the contrary, product 1 obtained
D
CHCl3)]. In fact, when comparing this optical rotation to the specific
optical rotation observed for enantiomerically pure compound the
ee of 1 corresponds to an only ee <60%.
Although, from a thermodynamical point of view these obser-
vations are in agreement with our previous reports concerning the
preferred formation of the amides in lipase-catalyzed reactions
carried out in polar medium,6,7 the highly selective accumulation
of the Michael adduct as a function of the nature of reaction
Ph
Me
NH
O
OMe
O
3
CaLB
+
+
O
Ph
NH2
Me
OCH
3 72 h, 65 °C
media, arises as
engineering.
a new and original application of solvent
1
2
Solvent
Me
N
H
Ph
A closer look at these results, in Figure 2, shows a clear corre-
lation between the Michael addition reaction selectivity and the
reaction medium polarity [as measured by the empirical parameter
ET(30)]. This confirmed that the rational selection of the solvent (or
solvent mixtures) directs the chemoselective process toward the
4
Scheme 1. Addition of benzylamine to methyl crotonate catalyzed by CaLB in organic
medium to yield Michael addition 3 or aminolysis product 4.
The hypothesis of this work is that rational changes in the po-
larity of the reaction medium can control the selective formation of
the Michael adduct 3 or the aminolysis product 4.
A
In the first step, the polarity of the reaction medium was
selected according to the empirical polarity parameter ET(30) de-
scribed by Reichardt8 for pure solvents and adapted by Castillo et al.
for solvent mixtures.6c Contrary to dielectric constants, which
rather reflect bulk polarities, ET(30) take into account solvent–solute
interactions at the molecular level.9
Seven media ranging from ET(30)¼30.9 (non-polar) to
ET(30)¼41.0 (polar) were screened in enzymatic reactions at
65 ꢀC with the amount of Michael adduct 3 and aminolysis
product
4 quantified by HPLC. While neither product was
formed in the absence of CaLB, a large range of ratio between 3
and 4 was obtained under enzyme-catalyzed condition which, as
expected, was strongly driven by the polarity of the reaction
medium.
From the results in Table 1, one can observe that in hydrophobic
medium such as n-hexane or toluene Michael adduct is the pre-
dominant product while the aminonolysis product is the preferred
one in more polar solvents. Accordingly, a variety of ratios of 3 and 4
can be reliably obtained through controlling of the polarity of sol-
vent mixtures.
B
Table 1
Influence of polarity of the enzymatic reaction medium on the chemoselectivity
between Michael addition and aminolysis
Entry Solvent
ET(30) Ratio (%)
3 (mmol/mL) 4 (mmol/mL)
3
4
1
2
3
4
5
6
7
n-Hexane
Toluene
Diisopropyl ether
THF
n-Hexane/2M2B, 1:1 39.2
n-Hexane/2M2B, 1:3 40.1
2M2B
30.9
33.9
34.0
37.4
95
71
40
30
35
26
20
5
0.40
0.25
0.11
0.09
0.14
0.09
0.08
0.02
0.10
0.17
0.20
0.25
0.26
0.30
29
60
70
65
74
80
41.0
Figure 2. Plots of % of product 3 and 4 as a function of the solvent polarity parameter
ET(30).
Reaction conditions: 0.5 M benzylamine 1, 0.75 M of methyl crotonate 2, and 10 mg/mL
of CaLB at 65 ꢀC for 72 h. 3 Michael addition and 4 aminolysis products.

538
J. Priego et al. / Tetrahedron 65 (2009) 536–539
Table 2
Table 3
Influence of the initial benzylamine 1 concentration in n-hexane on the ratio of
Michael adduct 3 to aminolysis products 4
Influence of the presence of 4 in the course of the reaction of benzylamine 1 and
methyl crotonate 2 using 10 mg/mL of CaLB at 65 ꢀC for 72 h
Entrya 1 (mmol/mL) Ratio (%)
Yield 1b (%) 3 (mmol/mL) 4 (mmol/mL)
S0 (mmol/mL)
Products (mmol/mL)
Ratio (%)
Yield (%)
1
3
4
1 and 2
4
3
4
3
4
1
2
3
4
0.50
0.25
0.10
0.05
>95
75
8
<5 85
0.400
0.130
0.003
d
0.025
0.042
0.036
0.039
0.5
0.020
0.420
0.033
93
7
90.6
25 68
92 81
100 79
S0¼initial concentration of substrates.
d
a
In all entries, a ratio of 1.5 mmol of methyl crotonate 2 by each mmol of ben-
adduct formation became extremely slow. Actually, product 3 is
only detected after 6 h of reaction (Fig. 3). Surprisingly, in all ex-
periments after 72 h of reaction, the concentration of 4 never
exceeded 0.04 M.
zylamine 1 were used.
Determined by HPLC and referred to the initial concentration of benzylamine 1.
Ratios also were determined by HPLC after 72 h of reaction.
b
Although these initial rate experiments describe a kinetic con-
trol on chemoselectivity, the effect is observed only at very low
substrate concentrations.
Michael addition or aminolysis and shows that solvent engineering
may be used as an effective tool to design or predict the lipase-
catalyzed chemoselectivity.
In order to further explore the thermodynamic effect of initial
substrates concentration on chemoselectivity, we decided to vary
the substrates concentration after it has reached its maximal
value, with the purpose to switch from aminolysis to Michael
adduct products. The experiment was started with low substrate
concentration (0.1 M of both substrates in n-hexane) and the
products measured after 72 h of reaction resulting in 0.039 M of 4
and 0.03 M of 3. At this point, the concentrations of both sub-
strates were raised to 0.4 M and again after 72 h of reaction the
products yield was measured. As expected, in this second step
only product 3 was formed once 4 reached its maximal concen-
tration (0.035 M).
In an apparent contradiction to these results, two previous
reports of reactions between amines and
a,b-unsaturated car-
boxyl groups in hydrophobic solvents at low concentrations of
amine (0.13 M), and high ratio of enzyme/amine (50 mg of CaLB/
mmol of benzylamine) resulted in the preferential formation of
the amide over the Michael addition product.11a,b As the main
difference with our results is the low substrate concentration
used by the authors, a series of experiments were performed in
n-hexane, varying the initial concentration of amine 1 [molar
ratio 1:1.5 with respect to 2] with 10 mg of enzyme/mL in all
experiments.
Interestingly, after 72 h of reaction, a maximal concentration
of the amide 4 was obtained in all cases (w0.035 M). However, in
reactions with low initial substrate concentration of amine 1
a strong decrease in the Michael adduct product 3 occurs. It is
clear that under these circumstances, chemoselectivity of the
process at initial amine concentration lower than 0.25 M is fa-
vored toward the synthesis of 4 (Table 2). In fact, in agreement to
Puertas et al. (1993),11a,c at amine concentrations lower than
A second experiment was also carried out, this time a reaction
medium with the purified product 4 at its maximal solubility in
n-hexane (0.02 M) and 0.5 M initial concentration of 1 and 2 was
prepared. Again, and as predicted from our previous results, after
72 h of reaction, 0.42 M of 3 was observed, while 4 increased only to
0.033 M (Table 3).
From these results, it is clear that at high initial substrate con-
centration, the maximal concentration of amide is independent of
the reaction rate and is only a function of thermodynamic condi-
tions. We have shown previously6,7 that maximum solubility of
polar products in hydrophobic solvents such as hexane, limit the
yield of product. Therefore, even if the amide synthesis is faster at
low substrate concentrations, once the maximum concentration is
reached, its formation ceases, while the formation of adduct 3
continues until the global thermodynamic equilibrium is reached.
In short, substrate concentration can also be considered as
a factor affecting chemoselectivity, but only at very low concen-
trations and in a limited range of chemoselectivity (Table 2).
Therefore, if high productivities of either aminolysis or Michael
adduct products are required, then the optimum conditions for the
reaction is best controlled via the solvent polarity.
0.1 M the reaction is practically selective for the amide
synthesis.
4
From a kinetic point of view, it has been reported that ami-
nolysis reactions are faster than the Michael adduct formation with
CaLB.6c However, the rate is influenced by the nature of the sub-
strates and their initial concentration. In experiments carried out in
n-hexane at a high initial substrate concentration (0.5 M), the initial
rate for the Michael adduct reaction was higher than that of ami-
nolysis. On the other hand, at a low initial amine concentration (0.1
and 0.05 M) the aminolysis reaction was faster while the Michael
3. Conclusions
We have demonstrated that solvent engineering is an effective
tool to control chemoselectivity in the reaction between methyl
crotonate 2 and benzylamine 1 catalyzed by CaLB, the results show
that the accumulation of the Michael adduct 3 is favored in more
hydrophobic media while the amide product 4 in more polar sol-
vents, as long as high substrates concentration are used in the
reaction.
In addition, an appropriate solvent or solvent mixture may be
designed according to the empirical polarity parameter ET(30) and
used as an effective tool to predict chemoselectivity. In addition,
these results open a wide spectrum of possibilities to rationally
modify reaction media in biocatalytic reactions controlled by
equilibrium.
Figure 3. Initial reaction experiments for the lipase-catalyzed synthesis of Michael
adduct and aminolysis products. Reaction at 0.5 M initial substrate concentration: >
Michael adduct 3 and A aminolysis product 4. Aminolysis product at :¼0.1 M and
C¼0.05 M.

J. Priego et al. / Tetrahedron 65 (2009) 536–539
539
4. Experimental
4.3.2. N-Benzylcrotonamide 4
White solid, mp 114–116 ꢀC (lit.11a 113–115 ꢀC), yield 76%,
0.488 g. Spectral data of 1H and 13C NMR are consistent with the
literature reports.
4.1. Materials and methods
candida antarctica lipase B, CaLB Novozyme 435, was obtained
´
from Novozyme-Mexico A/C. Benzylamine and methyl crotonate
Acknowledgements
were obtained from Aldrich. All the solvents were distilled over an
appropriate desiccant under nitrogen. Silica gel of 70–230 mesh of
Merck was used for purification by flash chromatography.
This research was supported by CONACyT (grant 48356-Q) and
by PAPIIT-UNAM (grants IN226706-3). J.P.-O. and M.C., C.O.-N.
thank CONACyT for a postdoctoral fellowship and graduate schol-
arships, respectively. We thank Fernando Gonzalez for technical
4.2. Analytical methods
´
assistance, Novozymes-Mexico for their generous donation of CaLB,
Spectra data of 1H and 13C NMR of Michael adduct 3 and ami-
nolysis products 4 as well as the proportions of each product in the
reaction mixture were recorded on a Varian Gemini 200 Spec-
trometer using CDCl3 as solvent and TMS as an internal standard.
The conversions were determined on an HPLC Waters 510 pump,
using a diodes array detector at 210 nm equipped with a C18 re-
versed phase Novapack column (3.9ꢁ75 mm). Substrates and
products were eluted with a mixture 80:20 of phosphate buffer
50 mM pH 2.6/MeOH at a flow rate of 1.0 mL minꢂ1, retention
times: 1¼0.86, 3¼1.49, 2¼3.70, and 4¼6.83 min.
and Dr. Minnhuy Ho for many important observations.
References and notes
1. (a) Bornscheuer, U. T.; Kazlauskas, R. J. Hydrolases in Organic Synthesis: Regio-
and Stereoselective Biotransformations; Wiley-VCH: Weinheim, 1999; (b) Carrea,
G.; Riva, S. Angew. Chem., Int. Ed. 2000, 39, 2226–2254; (c) Lutz, S. Tetrahedron:
Asymmetry 2004, 15, 2743–2748; (d) Flores-Sa´nchez, P.; Escalante, J.; Castillo, E.
Tetrahedron: Asymmetry 2005, 16, 629–634.
2. (a) Bertau, M. Curr. Org. Chem. 2002, 6, 987–1014; (b) Bornscheuer, U. T.; Bessler,
C.; Srinivas, R.; Krishna, S. H. Trends Biotechnol. 2002, 20, 433–437; (c) Shaw, N.
M.; Robins, K. T.; Kiener, A. Adv. Synth. Catal. 2003, 345, 425–435.
3. Torre, O.; Alfonso, I.; Gotor, V. Chem. Commun. 2004, 1724–1725.
4. (a) Cai, Y.; Yao, S. P.; Wu, Q.; Lin, X. F. Biotechnol. Lett. 2004, 26, 525–528; (b) Cai,
Y.; Wu, Q.; Xiao, Y. M.; Lv, D. S.; Lin, X. F. J. Biotechnol. 2006, 121, 330–337.
5. (a) Branneby, C.; Carlqvist, P.; Magnusson, A.; Hult, K.; Brinck, T.; Berglund, P.
J. Am. Chem. Soc. 2003, 125, 874–875; (b) Branneby, C.; Carlqvist, P.; Hult, K.;
Brinck, T.; Berglund, P. J. Mol. Catal. B: Enzym. 2004, 31, 123–128; (c) Carlqvist,
P.; Svedendahl, M.; Branneby, C.; Hult, K.; Brinck, T.; Berglund, P. ChemBioChem
2005, 6, 331–336; (d) Svedendahl, M.; Hult, K.; Berglund, P. J. Am. Chem. Soc.
2005, 127, 17988–17989.
4.3. Lipase-catalyzed Michael addition and aminolysis
The typical procedure consisted in the preparation of a solution
containing 0.39 g of benzylamine (3.67 mmol), 0.55 g (1.5 equiv,
5.51 mmol) of methyl crotonate, and 0.07 g of CaLB in 7.5 mL of
a pure dry solvent or a solvent mixture. The reaction medium was
heated to 65 ꢀC for 72 h in a 10 mL sealed vial. Afterward, the re-
action mixture was filtered and the enzyme washed with ethyl
acetate (3ꢁ3 mL). The filtrated solvents were evaporated under
reduced pressure to give crude yellow oil, which was analyzed by
HPLC to determine the composition of 3 and 4. Both products were
purified by silica gel flash chromatography with an n-hexane/ethyl
acetate gradient from 80:20 to 60:40 and their structures were
confirmed by 1H and 13C NMR.
6. (a) Bellot, J. C.; Choisnard, L.; Castillo, E.; Marty, A. Enzyme Microb. Technol. 2001,
28, 362–369; (b) Rendo´n, X.; Lo´ pez-Munguı´a, A.; Castillo, E. J. Am. Oil Chem. Soc.
´
´
2001, 78, 1061–1066; (c) Castillo, E.; Pezzotti, F.; Navarro, A.; Lopez-Munguıa, A.
J. Biotechnol. 2003, 102, 251–259.
´
´
´
7. (a) Reyes-Duarte, D.; Castillo, E.; Martınez, R.; Lopez-Munguıa, A. Biotechnol.
Lett. 2002, 24, 2057–2061; (b) Castillo, E.; Lo´ pez-Gonza´lez, I.; De Regil-Her-
na´ndez, R.; Reyes-Duarte, D.; Sa´nchez-Herrera, D.; Lo´pez-Munguı´a, A.; Darszon,
A. Biochem. Biophys. Res. Commun. 2007, 356, 424–430.
8. (a) Reichardt, C. Angew. Chem., Int. Ed. Engl. 1979, 18, 98–110; (b) Reichardt, C.
Chem. Rev. 1994, 94, 2319–2358.
9. Sinkeldam, R. W.; Tor, Y. Org. Biomol. Chem. 2007, 5, 2523–2528.
´
´
´
10. Torres-Gavilan, A.; Castillo, E.; Lopez-Munguıa, A. J. Mol. Catal. B: Enzym. 2006,
41, 136–140.
4.3.1. (ꢃ)-Methyl 3-(benzylamine)butanoate 3
20
11. (a) Puertas, S.; Brieva, R.; Rebolledo, F.; Gotor, V. Tetrahedron 1993, 49, 4007–
Yellow oil, yield 70%, 0.53 g, [
a
]
0.45 (c 1.0, CHCl3) for product
D
obtained in n-hexane, [a]
20 11.0 (c 1.0, CHCl3) for product obtained
´
4014; (b) Sanchez, V.; Rebolledo, F.; Gotor, V. Synlett 1994, 529–530; (c) Torre,
O.; Gotor-Ferna´ndez, V.; Alfonso, I.; Garcı´a-Alles, L. F.; Gotor, V. Adv. Synth. Catal.
2005, 347, 1007–1014.
D
in 2M2B. Spectral data of 1H and 13C NMR are consistent with the
literature reports.12
12. Asao, N.; Uyehara, T.; Yamamoto, Y. Tetrahedron 1988, 44, 4173–4180.